Flexible sensitive K-anonymization on transactions

In recent years, privacy breaches have been a great concern on the published data. Only removing one’s personal identification information is not sufficient to protect individual’s privacy. Privacy preservation technology for published data is devoted to preventing re-identification and retaining the useful information in published data. In this work, we propose a novel algorithm to deal with sensitive and quasi-identifier items, respectively, in transactional data. The proposed algorithm maintains at least the same or a stronger privacy level for transactional data with 1/k. In numerical experiments, our proposed algorithm shows better running time and better data utility.